Key messages

  • The acceptability of a digital intervention, combining text messaging service with a smartphone app, to support medication adherence in patients with hypertension in the UK primary care, has not been evaluated.

  • This pilot study proved that the digital intervention is acceptable among patients with hypertension, and it recommended that an interactive intervention to enable patients track health-related behaviours might be the way forward to provide personalised and highly tailored advice and support for adherence in primary care.

  • Future trial should investigate the feasibility of the digital intervention to support medication adherence and associative blood pressure in patients with hypertension, as an adjunct to usual care consultations.

Introduction

In England, over 12.5 million people are diagnosed with hypertension (high blood pressure) [1]. High blood pressure is a major risk factor for morbidity and mortality [2]. Taking anti-hypertensive medication as prescribed can significantly reduce these risks [3]; however, a substantial proportion of patients do not take their medication as prescribed. A recent meta-analysis showed that 41% of people do not adhere adequately to anti-hypertensive medications, i.e. they take less than 80% of their prescribed tablets [4]. Non-adherence contributes to increased hospital admissions, additional consultations, referrals, investigations, and medicine wastage. Improved adherence could save of just over £100 million per year; thus, the Department of Health recommends that novel interventions for medication adherence should be developed and tested [5].

Non-adherence to prescribed medication may occur for a number of reasons, like patients forgetting or missing a dose or a day of their medication. These reasons are described as non-intentional non-adherence (NINA) or intentional non-adherence (INA) [6]. Interventions are likely to be more effective to support medication adherence, when they address either or both of these determinants by providing highly tailored advice to the individual [7]. Tailored interventions include behaviour change techniques (BCTs) mapped onto these key determinants. BCTs are assumed to be the ‘active ingredients’ of behavioural interventions [8]. For example, INA can be addressed by reinforcing positive beliefs about taking medications (e.g. ‘keep control of your blood pressure today by taking all your tablets as prescribed’) and by countering negative beliefs and concerns (e.g. ‘please do not forget to take your tablets, even when you do not have any symptoms. Pills have been prescribed to lower your blood pressure’). NINA can be addressed through explicit and implicit reminders [9]. Both NINA and INA can be addressed by simple query messages to report behaviour (e.g. ‘have you taken your medications as prescribed during the past 7 days?’).

Digital interventions like text messaging and downloadable applications on smartphones (apps) are promising ways to provide advice, reminders, and encouragement, and to support patients to take their tablets as prescribed. Our recent meta-analyses suggested that automated telephone-based interventions, including text messaging interventions [7] and apps [10], double the odds for adherence to medication prescribed for long-term health conditions. Our recent feasibility trial found that digital interventions, including text messaging, can provide highly tailored advice to address either or both NINA and INA for medication adherence and is feasible and potentially effective adjunct to primary care consultations [11]. However, there is no intervention that has combined text messaging with a smartphone app, both to maximise the reach of the intervention and to utilise the advantages of the automated tracking technology embedded into smartphone apps when facilitating a highly tailored behaviour change intervention.

Based on our previous promising findings [11], we have developed a 56-day prototype intervention, which we call PAM (Programme on Adherence to Medication), delivered by a text messaging service followed by a smartphone app. This pilot study aimed to pre-test its acceptability to support medication adherence in patients with hypertension in primary care. To our knowledge, no other text messaging followed by a smartphone app intervention has been pre-tested to address non-adherence to medication in the UK primary care setting.

Qualitative and quantitative data was integrated into a mixed methods analysis to answer the primary research question of this pilot study: is an individually tailored text messaging service followed by a smartphone app acceptable to support medication adherence by patients treated for hypertension in primary care? Our secondary research questions were to explore the acceptability of the intervention content and delivery mode functionalities, and the mechanism by which these might have influenced medication-taking behaviour.

Methods

Recruitment methods and procedure

Three primary care practices were recruited in this study, which were located in East of England and at diverse deprivation areas. Patients were deemed eligible to participate in this study if they (a) had a diagnosis of hypertension or high blood pressure, (b) were prescribed at least one anti-hypertensive medication for a duration of at least 3 months before recruitment, (c) be had poorly controlled hypertension (e.g. readings of blood pressure > 140/90 mmHg recorded in practice databases during a period of 6 months before recruitment), (d) were aged 18 years or older, (e) had a good understanding of English, (f) owned and regularly used a mobile phone, and (g) had the capacity to provide informed consent. We aimed to recruit 25 participants to respond to our research questions.

Across the three recruited practices, a total of 1340 patients eligible to take part in this study were identified from practice records. A member of the practice staff sent one text message invitation to all eligible patients. The invitation included a link to access the study material, and to register interest to participate, online. In total, 79 patients responded to the text message invitation with an interest to participate during the first week from the day invitations were sent, 54 of whom in the first 2 days. From those, the first 23 patients who met all the eligibility criteria to participate were enrolled in the intervention during a baseline meeting with a member of the research team. During the baseline meeting, a member of the research team obtained written informed consent and facilitated the collection of baseline data. Two researchers conducted the recruitment meetings from January until March 2019.

Intervention

Participants were provided with the 56-day prototype intervention, which involved text messaging support for 28 days and then switching to the smartphone app for the consecutive 28 days. All patients were asked whether they would like to switch from the text messaging to the app intervention, after they have completed the text messaging intervention. The app intervention was compatible with android phones only. 

Participants received daily reminder messages with explicit advice (e.g. ‘please do not forget to take your medication today: amlodipine, 2 tablets, 5mg’), daily advice messages (e.g. ‘the health benefits of taking your meds regularly is having low blood pressure. Please keep looking after yourself by taking your medications’), and query messages (e.g. ‘have you taken all your prescribed meds in the last 7 days/ today? Reply Yes or NO’). The app included additional functions; provided participants with an option to request and receive feedback on behaviour and to change the delivery of the intervention messages. The app tracked medication related routines without feedback to the patient.

Data collection methods and procedure

Data were collected using questionnaires, telephone interviews, and digital log files.

Questionnaires

The questionnaire included questions regarding patients’ intention to adherence and beliefs about taking medication. Intentional (INA) and non-intentional non-adherence (NINA) was measured using the 5-item MARS questionnaire [12], and beliefs about adherence were measures using 10 items measuring necessity beliefs (e.g. ‘If I were to take all my medications as prescribed without missing a day it would do more harm than good’), control beliefs (e.g. ‘taking my medications as prescribed keeps my blood pressure under control’), affective attitudes (e.g. ‘taking my medications as prescribed every day without missing a day is pleasant/unpleasant/neither’), social norms (e.g. ‘If they were prescribed tablets, most people whose opinion I value, would take all their prescribed tablets without missing a day’), copping self-efficacy (e.g. ‘I take all my medications as prescribed without missing a day, even if I am busy at home’), and generic emotional state (e.g. ‘How much of the time during the past month have you felt calm and peaceful’). The tailoring questionnaires were adapted from, and developed based on our previous studies [11, 13]. Patients’ answers to these questions were used to tailor the content of the advice messages.

The questionnaire at baseline collected information about patients’ prescribed medications (e.g. name of medications and dosage), and about the delivery of the intervention i.e. preferred time to receive the intervention messages. Patients’ answers to these questions were used to tailor the content of the reminder messages.

Patients completed questionnaires at baseline and at follow up. Baseline questionnaire collected information about patients’ demographic characteristics. Follow up questionnaires obtained views about the acceptability of the intervention. Baseline questionnaires were completed by patients during the meetings with a member of the research team and follow-up questionnaires were complete using an online webpage, which was emailed to patients.

Interviews

Semi-structured interviews were chosen to obtain in-depth accounts of participant experiences, thoughts, and beliefs. Detailed answers were provoked by the researcher, who used open-ended questions and prompts to explore participants’ experiences with using the intervention [14].

During the interviews, a member of the research team used a semi-structured interview guide to prompt patients’ views about the intervention content and delivery modes, and obtain recommendations for improvement (please see Appendix 1 and Appendix 2 for the semi-structured interview guide). Each participant was provided with the option of a weekly 15-min telephone interviews or one follow up 45-min telephone interview. Seventeen patients opted for the weekly calls, and the other five for the interview at the end of the trial. Interviews were audio recorded, and the audio files were transcribed by a third-party company. Patients who could not attend the interviews (n = 1) provided their feedback using emails. Interviews were conducted from February until April 2019.

Weekly data collection was selected to capture participant thoughts and opinions within its present context as opposed to potentially weeks later, thus deriving more detailed, richer data from each participant. Regular data collections also aimed to identify and solve any technical issues that may arise when piloting an innovative digital intervention.

Digital log files

The engagement with the intervention was captured by digital log files. Digital log files captured the intervention content participants received and interacted with (e.g. reported whether or not the medication was taken, check feedback on adherence to medication, used app settings to tailor message delivery). Additionally, digital log files captured medication related routines. The log files are documents detailing each participants’ recorded actions and responses whilst using the intervention and were utilised as objective measure of intervention usage and engagement. Data from log files were extracted by a member of the research team.

Analysis

Two members of the research team used thematic analysis and analysed all interview transcripts independently. The coding for emerging themes was deductive and aimed to provide answers to our research questions. Inductive themes were considered as recommendations for improvement. The two researchers compared notes and codes for each interview transcript and made a mind map to visually link quotes and codes to one another. The two researchers discussed in-depth themes that had emerged from the data. Nvivo software was used to facilitate data analysis.

Data from log files were coded by a member of the research team and integrated into the analysis. Data obtained from questionnaires and digital log files were summarised using descriptive statistics. A mixed method approach was used to synthesise quantitative and qualitative data to provide answers to our research questions [15].

Results

Seventy-nine patients answered to the text message invitation and expressed their interest to participate in the study (5.9% response rate to invitation), of whom the first 23 patients attended baseline meetings and provided written informed consent to participate. All 23 initially recruited patients completed the baseline questionnaire (100% completion rate), which included the tailoring questions and measured medication adherence. One patient dropped out of the study before registering with the intervention due to personal reasons. In total, 22 participants enrolled in the intervention: all registered with the intervention for 28 days using the text messaging service, four of them (17%) selected to switch to the app, and two installed the app and used the intervention for 56 consecutive days (see Fig. 1).

Fig. 1
figure 1

Study flow chart

Full size image

Twenty patients (out of the final 22 participants) completed the follow-up interviews (90.9% response rate), and 19 completed the follow-up questionnaires (86% completion rate). Participants were 64% men and 82% above 60 years old (see Table 1).

Table 1 Practice (n = 3) and participants’ (n = 22) characteristics
Full size table

The data analysis generated results in three main themes to provide answers to our research questions:

  1. 1.

    Acceptability and usage of the components (i.e. content and delivery functionalities) of the prototype digital intervention to patients prescribed medications to treat high blood pressure

  2. 2.

    Mechanisms by which these components might have supported medication adherence

  3. 3.

    Recommendations to improve the acceptability of the intervention

Acceptability and usage of the digital intervention in patients with hypertension

All participants interviewed (n = 20) reported that the text messaging service was easy to use (see Table 2 quote 1 and Fig. 3). Similarly, patients who used the app reported the app functionalities were easy to navigate (see Table 2 quote 2 and Fig. 3).

Table 2 Thematic analysis results
Full size table

Reminder messages

Commenting on the reminder messages, one participant thought the simple message sufficed and also allowed for the participant to ‘double-check’ that they had taken their medication that day (see Table 2 quote 3). Most participants reported the reminder messages as acceptable and a positive aspect of the digital intervention (see Figs. 2 and 3). However, few participants reported disliking the frequency of the daily reminder messages, with one participant finding them overbearing. To overcome this challenge, participants recommended that the intervention should include more options for participants to reduce the frequency of receiving these messages, especially if the intervention was of a longer duration (see Table 2 quote 4). Both app patients have used the snooze functionality of the app to re-schedule the time of the reminder messages (see Table 3).

Fig. 2
figure 2

Satisfaction with the intervention. N = 19 patients. Data collected by follow-up questionnaires

Full size image
Fig. 3
figure 3

Acceptability of the intervention as an adjunct to usual care. N = 19 patients. Data collected by follow-up questionnaires

Full size image
Table 3 Intervention engagement per delivery mode
Full size table

Advice messages

Participant opinions on the advice, non-reminder, messages were much more variable (see Fig. 2). For example, one participant suggested that the advice messages increased his attention to read the messages and in turn increased his motivation to take his medication as suggested (see Table 2 quote 5). This may be due to receiving a variety of advice messages throughout the digital intervention, with each message emphasising and addressing a different reason regarding medication adherence. Data from digital log files suggested that participants who used the app acknowledged or dismissed the receipt of all advice messages.

Query messages

Furthermore, the ‘query’ message was sent to participants asking them to reply ‘yes’ or ‘no’ as to whether they had taken all their prescribed medication. Participants were asked to respond to the query message one per week (when testing the text messages) or when each medication was taken (when switching to the smartphone app). The concept of the query message arose from the behaviour change strategy of ‘reporting whether or not the behaviour was performed’ [7, 11, 16]. All participants reported to query messages and most reported that they liked these messages (see Table 3 and Fig. 2). All participants received the query messages and the response rate, to both text message and app notifications, was 100%, and this engagement score was maintained throughout the intervention, indicating high participant engagement with the digital intervention and acceptability of these messages to support medication adherence.

Participant discussed the benefits of the ‘feedback on behaviour’ functionality on the app, which generates a percentage score of patients’ self-reported medication adherence over the past day and week (e.g. ‘you’ve reported that you have taken 89% of your prescribed medication last week’). Participants found this intervention content informative and reported using it to self-monitor their medication adherence and to reflect on their patterns regarding their medication-taking routine (see Table 2 quote code 6). Both app users requested the feedback on their self-reported behaviour by checking the app functionality, in most cases after they have self-reported medication adherent behaviour (see Table 3).

Functions to tailor message delivery

This function provides the participant with options to re-schedule the frequency (e.g. by texting 'less' to receive less messages or 'more' to receive more intervention messages) and the time of the reminder messages (e.g. by snoozing the reminder notifications). Participants tailored the delivery of the reminder messages on average 5 times when using the smartphone app, and none when using the text messages. This function was found useful to those participants who reported that were busy at the time the reminder message appeared and could not respond confirming whether or not they have taken their tablets. Patients have also the option to set and re-schedule the time and date of their refill prescription reminders, and both patients set their refill reminders and used the app function to re-schedule the time and day of the refill reminder notification on average 7 times, during 4 weeks (see Table 3). In all occasions, they tailored the delivery of the reminder messages successfully, implying the importance of including this function in the app, as well as how easy it is to use it. Some participants suggested that the option to tailor the delivery of the refill reminder notifications increased the acceptability of the intervention.

Acceptability of tracking medication related routines using the smartphone technology

All participants were asked about their views about the app using the GPS, Wi-Fi, and accelerometer embedded on their phone, to automatically track their behaviours related to medication taking during the intervention. All participants reported having no concerns over the security of their behavioural patterns related to medication adherence being captured through the app sensing technology. Some of them have also suggested that tracking technology could make daily medication taking effortless and easier (see Table 2 quotes 7 and 8). They also suggested that the information about their medication taking routines, collected by smartphone technology, could be utilised during consultations with health care providers to improve medication adherence.

Mechanisms by which the intervention has supported medication adherence

For some participants, receiving the query message was an opportunity to reflect on their medication-taking behaviour and raise their awareness as to whether or not they had taken all the doses of the medication they were prescribed (see Table 2, quotes 9, 10, and 11, and Fig. 3). This strategy seemed to have motivated medication taking through several ways: (a) by increasing participants’ commitment to reply to the query message, (b) by raising awareness of tablet-taking routine, (c) by increasing feelings of involvement with their own medication-taking routine, and (d) by empowering them to take their medications as prescribed (see Table 2 quotes 12, 13, and 14). These findings suggested that reporting behaviour is a motivational behaviour change technique, which can lead to increased performance of the target behaviour.

The importance of accessing feedback regarding medication adherence behaviour in motivating participants to change their behaviour, was also suggested by participants testing the app. Both app testers checked this app functionality after successfully installing the app (see Table 3), with one of the participants attempting to view their adherence report six times in the first day of switching to the app. This could be explained by an initial exploration period after first installing the app; however, both participants continued to check the functionality with feedback on behaviour at least once a day for 4 days after switching to the app. These findings imply that participants find this function interesting and an engaging feature of the app and that feedback on the behaviour might be important to increase patients’ motivation to change behaviour.

It was also found that the tailored daily advice messages may have caused an increase in the participants’ attention and curiosity when interacting with the intervention, and therefore an increase in engagement with the intervention. It was also recommended that the tailored advice messages have prompted self-monitoring and possibly increased acting upon these advice messages.

A number of participants agreed that the digital intervention may also help to increase patients’ feeling of empowerment and ownership of their long-term health condition, with one participant expressing how he particularly liked one of the advice messages which encouraged patients to take responsibility for their own medication-taking routines as a self-care process. Participants reported that they found the intervention messages particularly useful because they suported them to establish a sense of ownership over the condition and motivated them to be more self-aware of their independent responsibility over their health condition (see Table 2 quotes 15, 16, 17, and 18).

All participants liked the interactive elements of the intervention, and reported feeling motivated to continue taking tablets regularly and as prescribed, after interacting with the intervention (see Table 2 quote 15).

Most participants expressed preferences towards the advice messages which included positive reinforcement to take their tablets (see Table 2 quotes 17, 18, and 19). The preference towards the advice messages came up multiple times during the interview, specifically in the context of increasing patients’ motivation to take their medication to treat hypertension as a self-care process (see Table 2 quotes 15 and 19).

Recommendations to improve the acceptability of the digital intervention

All patients reported that they would recommend the intervention to other people who have been prescribed medications for long-term health conditions and those newly prescribed (see Fig. 3). They have also made recommendations to improve intervention content, delivery, and implementation procedures.

Many participants suggested receiving separate reminder messages for every medication they take each day. However, many other participants reported patients who are prescribed multiple daily medications a potential limitation to this idea, for example sending individual reminder messages to an individual who takes ten or more tablets every day will most probably be less acceptable (See Table 2 quote 20 and 21). However, another participant recommended a solution to this problem, suggesting combining the medication reminders into morning, lunchtime, and evening routines (see Table 2 quote 22).

Furthermore, one of the participants testing the app reported not using the additional app functions, which could be used to tailor the delivery of the medication reminder until a later time (see Table 2 quote 23). The participant recommended a help guide with interactive elements to explain all features of the app and how to use them when first installing the app (see Table 2 quote 24 and 25). A help guide to aid patients’ navigation through the app would potentially increase accessibility and intervention engagement.

Participants also recommended different ways to summarise the feedback on behaviour, such as graphs or colour-coded systems and traffic light colour code systems (green = goal met; orange = room for improvement; red = goal unmet, improvement needed), which are universally understood and easy to interpret, thus might be accessible to a range of individuals with varying learning styles (see Table 2 quote 26).

Moreover, the frequency of the query message was also discussed between participants; some participants suggested that responding to the query message every day would be effortless and thought others would prefer less frequent queries about medication adherence. However, another participant suggested receiving a daily query message, stating that it would help those with poor memory or those with busy daily routines (see Table 2 quote 27).

Towards the end of the 28-day intervention, many participants suggested once a week, instead of every day, was an appropriate frequency to receive the advice messages to improve intervention acceptability and engagement (see Table 2, quote 28). However, one participant expanded upon this idea and suggested that the impact of the advice messages might be decreased if received daily (see Table 2 quote 29). Thus, receiving one or two advice messages a week would be more acceptable to encourage patients to take their prescribed medications regularly, especially if the intervention was to support sustained medication adherence.

The tracking functions of the digital intervention could facilitate effortless medication adherence in the long term, however to increase its acceptability, a clear and honest communication is needed to explain the privacy and the purpose of using tracking data to support medication adherence (see Table 2 quote 30). Furthermore, some participants also suggested using the logo of the NHS or University of Cambridge to visually link the app to a trusted institution (see Table 2 quote 31).

Overall, participants suggested that the digital intervention could support medication taking daily routines and be an acceptable adjunct to primary care and increase satisfaction with the health care provided by GP practices (see Table 2 quotes 32 and 33 and Fig. 3).

Discussion

This mixed methods study suggests that the PAM prototype intervention is acceptable by patients with hypertension and that it may support medication adherence by improving motivation to take all doses of the prescribed medications as part of a daily routine. It was found that the content of the highly tailored intervention was acceptable and that all participants used and engaged with the intervention. The study also supported that all patients interacted with the query intervention messages to support their medication-taking behaviour and they found the additional tracking functionalities of the smartphone technology acceptable, with few (17%) opting to use these additional tracking elements to support medication adherence.

The high response rate to the messages enquiring participant to report about adherence to medication demonstrates a promising projection of engagement that will be especially important for the future trial of this intervention, to provide information about how intervention engagement associates with changes in behavioural and clinical outcomes. Thus, not only do the results of this study suggest that both the text messaging and app delivery modes of the digital intervention are acceptable, but also that the content of the intervention is acceptable by patients with high blood pressure. Therefore, the positive results from this study provide us with confidence to proceed into testing this digital intervention for feasibility at a randomised controlled trial.

A strength of this pilot study is that it utilised two digital delivery modes to maximise the reach of a highly tailored intervention even to those with low digital literacy and to provide with a choice of an automated self-tracking advice to support medication adherence. To our knowledge, this is the first intervention that utilised a combination of text messaging and smartphone app and has been pre-tested to address non-adherence to medication within the UK primary care setting. Another strength of this study was its unique data collection method of weekly interviews that allowed us to identify and solve any technical issues more explicitly and quickly than using the traditional hour-long interview at the end of the trial. This study also used a mixed methods analysis and synthesised qualitative and quantitative data from multiple sources to respond to our research questions.

This study is limited by the small number of patients as well as the short duration of the intervention. Another limitation of the study is the small proportion of patients switching to the smartphone app.

Conclusion

The findings of this study informed the upcoming trial to evaluate the feasibility of this digital intervention to support medication adherence and blood pressure to patients prescribed treatement for hypertension in primary care. This study provides evidence that this novel digital intervention is acceptable by patients with hypertension in primary care. Not only is the content of the intervention acceptable, but also the two digital delivery modes were found to be acceptable by patients. This study also found the digital intervention to be highly engaging and supportive to patients. When considering the high cost medication non-adherence imposes upon the NHS, this study is of particular importance as it provides evidence supporting that this low-cost intervention may be an acceptable answer to help achieve healthcare priorities.